System and Method for Call Classification

1. A method for analyzing and classifying a voice call, the method comprising: intercepting a voice call session between an initiating device and a recipient device; transforming voice call data exchanged between the initiating device and the recipient device during the voice call session into a predefined data format, wherein transforming the voice call data further comprises recording the voice call data and wherein the predefined data format comprises formatted text; analyzing the transformed voice call data to determine one or more attributes of the intercepted voice call; identifying one or more features associated with the intercepted voice call session based on the determined one or more attributes; and classifying the intercepted voice call using the one or more identified features, wherein classification type of the intercepted voice call includes at least one of: a regular call and unwanted call and wherein the one or more identified features include at least one of (i) presence of robotized speech and (ii) emotional component of the intercepted voice call, wherein the presence of the robotized speech is determined using a set of previously selected strings of phonemes and wherein the emotional component of the intercepted voice call is positive or negative based on at least one of: (i) rules in which text is broken up into words or sequences of words having a previously assigned positive or negative evaluation, (ii) glossaries in which the number of positive and negative words from a previously compiled glossary are counted, (iii) machine learning.

2. The method of claim 1 , wherein the call category includes at least one of: telemarketing, social survey, offer of services, fraud.

3. The method of claim 2 , wherein the call category is determined using one or more call categorization attributes based on the recording, and wherein the call categorization attributes include at least one of: words, n-grams, word-embedding, bag-of-words.

4. The method of claim 1 , wherein classifying the intercepted voice call further comprises using a trained classification model based on a set of previously selected voice calls having features.

5. The method of claim 4 , wherein the trained classification model is based on at least one of: naive Bayesian classifier, logistic regression, MRF (Markov Random Field) classifier, SVM (support vector machine), k-nearest neighbor, and/or decision tree.

6. The method of claim 1 , further comprising sending a notification to the recipient device, wherein the notification indicates a classification category of the intercepted voice call.

7. A system for analyzing and classifying a voice call, the system comprising: a hardware processor configured to: intercept a voice call session between an initiating device and a recipient device; transform voice call data exchanged between the initiating device and the recipient device during the voice call session into a predefined data format, wherein the hardware processor configured to transform the voice call data is further configured to record the voice call data and wherein the predefined data format comprises formatted text; analyze the transformed voice call data to determine one or more attributes of the intercepted voice call; identify one or more features associated with the intercepted voice call session based on the determined one or more attributes; and classify the intercepted voice call using the one or more identified features, wherein classification type of the intercepted voice call includes at least one of: a regular call and unwanted call and wherein the one or more identified features include at least one of (i) presence of robotized speech and (ii) emotional component of the intercepted voice call, wherein the presence of the robotized speech is determined using a set of previously selected strings of phonemes and wherein the emotional component of the intercepted voice call is positive or negative based on at least one of: (i) rules in which text is broken up into words or sequences of words having a previously assigned positive or negative evaluation, (ii) glossaries in which the number of positive and negative words from a previously compiled glossary are counted, (iii) machine learning.

8. The system of claim 7 , wherein the call category includes at least one of: telemarketing, social survey, offer of services, fraud.

9. The system of claim 8 , wherein the call category is determined using one or more call categorization attributes based on the recording, and wherein the call categorization attributes include at least one of: words, n-grams, word-embedding, bag-of-words.

10. The system of claim 7 , wherein the hardware processor configured to classify the intercepted voice call is further configured to use a trained classification model based on a set of previously selected voice calls having features.

11. The system of claim 10 , wherein the trained classification model is based on at least one of: naive Bayesian classifier, logistic regression, MRF (Markov Random Field) classifier, SVM (support vector machine), k-nearest neighbor, and/or decision tree.

12. A non-transitory computer readable medium storing thereon computer executable instructions for analyzing and classifying a voice call, including instructions for: intercepting a voice call session between an initiating device and a recipient device; transforming voice call data exchanged between the initiating device and the recipient device during the voice call session into a predefined data format, wherein transforming the voice call data further comprises recording the voice call data and wherein the predefined data format comprises formatted text; analyzing the transformed voice call data to determine one or more attributes of the intercepted voice call; identifying one or more features associated with the intercepted voice call session based on the determined one or more attributes; and classifying the intercepted voice call using the one or more identified features, wherein classification type of the intercepted voice call includes at least one of: a regular call and unwanted call and wherein the one or more identified features include at least one of (i) presence of robotized speech and (ii) emotional component of the intercepted voice call, wherein the presence of the robotized speech is determined using a set of previously selected strings of phonemes and wherein the emotional component of the intercepted voice call is positive or negative based on at least one of: (i) rules in which text is broken up into words or sequences of words having a previously assigned positive or negative evaluation, (ii) glossaries in which the number of positive and negative words from a previously compiled glossary are counted, (iii) machine learning.